Benign soft tissue tumors (STT) and their malignant variants, sarcomas, form a highly heterogeneous group of tumors. To date, most gene expression profiling studies have examined relatively few cases each of several sarcoma types. The best hope for adequate treatment of patients suffering from malignancies lies in the continuing individualization of therapy for a particular patient. The molecular analysis of carcinomas has helped in identifying subgroups of patients that benefit more from one therapy over another. Perhaps the best example of this is found in breast carcinoma, where ER and HER2Neu status were found to be of great significance. More recently, gene array findings highlighting novel subtypes, such as the "basal subtype" give hope that further individualization of therapy is possible. No clear progress in this direction has been made in the field of sarcomas yet and tumors like leiomyosarcoma (LMS) and gastrointestinal stromal tumor (GIST) are treated as amorphous groups of lesions, despite the fact that LMS behaves in a manner that is influenced by the site of origin of this tumor and the fact that different mutations are known to exist in the KIT and PDGFRa genes in GIST that influence response to Imatinib. The objective of this project is to examine many (>100) samples from each tumor type to reach a molecular characterization of the different subsets of tumors that exist within these 2 categories. During the course of this project novel diagnostic and prognostic markers and novel potential therapeutic targets will be identified. Three relatively new technical developments will be used to achieve these goals: gene expression profiling combined with array-based comparative genomic hybridization to identify novel genes of clinical interest, followed by validation and extension of these findings on tissue microarrays analyzed by immunohistochemistry and/or in situ hybridization.